23

After watching Elon Musk's presentation regarding the Tesla Powerwall, I had a question.

On his diagram of the USA, he claims that a "tiny blue square" on the top of Texas that consists of solar farms would be enough energy to power the united states. I've watched documentaries, that claim it should be 10 times that size. This seems absurd right?

That blue square there is the land area that's needed to transition the United States to a zero-carbon electricity situation

enter image description here

Sklivvz
  • 78,578
  • 29
  • 321
  • 428
Outdated Computer Tech
  • 402
  • 1
  • 3
  • 11
  • 8
    Keep in mind that Musk never said that this area is needed to power the entire US - it is the area needed to remove fossil fuels. You would still be using hydro, nuclear and wind power, in addition to solar power. – T. Sar May 07 '15 at 19:28
  • 2
    It's on the first moments of his speech, right on when he tells about the blue square. He never talks about going full solar. I will do that in a moment. Just let me pick the correct second. – T. Sar May 07 '15 at 19:52
  • 4
    @EnergyNumbers It's on the 3:24 mark: "The blue square there is the land area that's needed to transition the united states to a zero carbon electricity situation". Zero carbon is different from Solar-only - Nuclear Power, Hydro Power and Wind power are all "zero-carbon", since they don't burn fossil fuels. – T. Sar May 07 '15 at 19:55
  • @Sickest have the edits to your question accurately captured the claim you wanted examining? This question [is being discussed on meta](http://meta.skeptics.stackexchange.com/q/3195/6236): I don't want to change the question to mean something you didn't intend, so if you can let us know if the question still reflects your intent or not, that would be great. – 410 gone May 08 '15 at 08:02
  • just to point out if it was all in one area like this the lost energy from transporting that power across all the US would lead to far larger amount of energy needed and thus more land used. I would also wish you good luck with making sufficient solar panels to cover an area this large ;) In short, whatever the theory is in practice it's clearly not possible. – dsollen Sep 01 '15 at 22:01
  • The discrepancy could come from the fact that "installed" solar cells are going to take up much more room than a solid sheet of cells that cover 10,000 sq km. Based on the math below from @EnergyNumbers, Musk seems to be using the "one giant sheet" method. You obviously can't pack in cells that densely over that much space. Needing 10x that space for a working installation isn't unreasonable. – JPhi1618 Sep 10 '15 at 19:14
  • 1
    @JPhi1618 you could pack it at 100% fill, but almost no one does for utility-scale, because it's not the most economical solution (except on rooftops). A 1:3 ratio is more likely for a typical commercial installation, but it will of course depend on cabling costs, land costs, location, and installation & maintenance techniques – 410 gone Sep 10 '15 at 19:22
  • @EnergyNumbers I was thinking about installations like [this](https://upload.wikimedia.org/wikipedia/commons/4/45/Giant_photovoltaic_array.jpg) when thinking about a 10x increase in size, but then I saw some pictures like [this](http://esarenewables.com/wordpress/wp-content/uploads/2011/12/Fuguay-Varina-solar-farm_0326-1024x768.jpg) that do seem to be MUCH more dense. – JPhi1618 Sep 11 '15 at 13:40
  • @JPhi1618 yes, your first image is of a *tracker* installation: because of the shadowing issue, trackers are installed at much lower density than fixed-mount panels (your second image). – 410 gone Sep 11 '15 at 14:30
  • The reason you may have seen areas 10x bigger are that many of these comparisons include spacing between their solar panels. This spacing allows maintenance/replacement crews to get to all the panels, and allows the use of solar panels that track the sun, which are much more cost effective than horizontal surface panels, but they do need spacing to avoid shading each other. And if all we cared about was to power the USA from solar with as little space as possible (and who cares about the expense of cleaning the solar panels using a helicopter), we could do that. – Scott Jun 05 '16 at 22:28
  • You could clean them without a helicopter--put the panels on hinges and mount them a bit above head height. Access is from underneath, unlock them and swing them down to clean or repair. The issue with flat panels is the increased cost of not tracking the sun. – Loren Pechtel Jun 06 '16 at 01:49
  • 1
    That tiny blue square is still VERY big. If you want to get to the centre of it to do some maintenance, you need a car. You need to connect your solar panels to HV lines to transport the power, and these will shade large areas. Sure, these wires COULD be underground if reducing land area is the goal, but it isn't, reducing cost is. There are a lot of reasons that space on a solar farm of this size would need to be bigger than the size of the panels. – Scott Jun 06 '16 at 02:28
  • I fully agree that you wouldn't actually build them that way, I was just saying that you wouldn't need to resort to a helicopter to clean them. – Loren Pechtel Jun 06 '16 at 02:29
  • Fair enough. Feel free to assume the helicopter comment was slightly tongue in cheek in the first place, and what I really meant was "there would be difficulties (and associated expenses) in cleaning, repairing and replacing panels unless appropriate spacing is left for access vehicles." – Scott Jun 06 '16 at 02:46
  • Hinging them down would address repairing and replacing also. It's just more expensive to build a structure like that than to lay them flat. – Loren Pechtel Jun 06 '16 at 02:47
  • The thing is, buying a chunk of land that is otherwise unusable will likely be more cost effective than most ways of shrinking down the footprint of the plant. Some analyses (such as this one) track the smallest amount of size that would be required, and some track how much a realistic solar farm would take up if people built it as cost effectively as they can, which explains the 10x difference the OP mentioned – Scott Jun 06 '16 at 03:15

2 Answers2

37

Summary: yes, the area shown is reasonable, as a visualisation of the surface area of panels required to generate electricity equal to total US electricity consumption, on a multi-year average: that area of panels would generate 500 GW, which is above the current US annual average electricity consumption of 425 GW

Calculations below are taken from this blogpost at the Energy Institute, University College London.

As Thales Pereira has pointed out in the comments on the question, the claim could be taken to mean that that area of PV would displace fossil fuel, and supplement the existing zero-carbon options on the US grid (hydro, nuclear, wind). That would make it an slightly weaker claim. I've assumed that it's the strong version of the claim: that that area would be the same amount of power as total US demand, even before considering existing hydro, nuclear and wind generation. If the stronger claim is valid, then the weaker claim automatically follows.

  • US electricity consumption is about 425 GW on average
  • The area shown is 10,000 km2, in NW Texas
    Eyeballing this google map and comparing with the graphic in the question: using the scale in the bottom-right corner, the square is about 100km along the side
  • Table 2 of Green et al's Solar cell efficiency tables (Version 45) gives the best module as being 24% efficient, giving a PV capacity per unit area of 0.24 GW/km2
  • PVWatts gives a capacity factor for Amarillo, Texas, of 21%

10,000 km2 x 0.24 GW/km2 x 21% = 500 GW

Which is more than current US electricity consumption of 425 GW.

So yes, the area shown is reasonable, as a visualisation of the surface area of panels required to generate electricity equal to total US electricity consumption, on a multi-year average.

410 gone
  • 4,701
  • 30
  • 45
  • Is solar thermal energy (such as [California Solar One](http://en.wikipedia.org/wiki/The_Solar_Project#Solar_One) or [Nevada Solar One](http://en.wikipedia.org/wiki/Nevada_Solar_One)) more or less efficient than PV panels? – vartec May 07 '15 at 21:00
  • 2
    @vartec It would be worth asking that question on [sustainability.se] There's little data to go on that I could find: AFAICT it looks like those (CSP) plants are less **land**-efficient than PV. They may have similar **energy** efficiency (~20%, before storage losses). **Cost**-efficiency is harder to compare: CSP's potential storage will have different values depending on what the rest of the grid looks like; and we don't really have enough experience to calculate a meaningful learning curve for CSP capital costs. – 410 gone May 08 '15 at 02:03
  • For the sake of this question I guess only land-efficiency matters. – vartec May 08 '15 at 16:43
1

Firstly, the northern 55 km of the square is the Oklahoma Panhandle, while the remainder is the Texas Panhandle, which is 270km east to west.

The blue square is therefore about 123km on each side, 15,000 sq. km. or 3,700,000 acres.

A great resource for evaluating whether this is sufficient is Land-Use Requirements for Solar Power Plants in the United States by the National Renewable Energy Lab.

According to table 4 (which considers only direct land use) the most efficient use is by 2-axis concentrating PV systems, which need only 2 acres to produce a GWh per year.

So with 3,700,000 acres, 1,850,000 GWh can be produced per year.

This compares to 4,686,400 GWh per year in electricity consumption for the USA as of 2013.

So, no, even considering the best current technology and only direct land use, it is not a large enough area.

DavePhD
  • 103,432
  • 24
  • 436
  • 464
  • This answer does not answer the question - 'electricity consumption for USA' != energy needed to transition to zero carbon electricity. – NPSF3000 Jun 06 '16 at 22:14
  • @NPSF3000 67% is currently from fossil fuels http://www.eia.gov/tools/faqs/faq.cfm?id=427&t=3 – DavePhD Jun 07 '16 at 00:42
  • so taking that into account, the difference between 4,686,400GWh*67% and 1,850,000GWh within a factor of 2 (70%), you may want to change answer to 'possibly' or firm up the math. – NPSF3000 Jun 07 '16 at 01:03
  • For example estimates for 2016 are only 4,088,000 GWh of electricity will be generated, 64.5% of which will be gas and coal - significantly different than the numbers you use. https://www.eia.gov/forecasts/steo/report/electricity.cfm – NPSF3000 Jun 07 '16 at 01:09
  • @NPSF3000 maybe the Wikipedia number is wrong. I'll investigate more. – DavePhD Jun 07 '16 at 02:03
  • @NPSF3000 It depends if you consider gross generation, net generation, or end use https://www.eia.gov/totalenergy/data/monthly/pdf/flow/electricity.pdf Gross is more like 4,330,000 GWh – DavePhD Jun 07 '16 at 11:39